A current protection setting optimization strategy based on a constraint-containing multi-objective particle swarm algorithm

Abstract

The application discloses a current protection setting optimization method based on a constrained multi-objective particle swarm, and belongs to the technical field of relay protection of power systems. The application comprises the following steps: (1) according to the 'four-property' requirement of relay protection, quantifying the optimization indexes of relay protection, and establishing a multi-objective optimization model of relay protection setting; (2) determining the optimization variables of current protection setting; (3) setting the constraint conditions of the setting variables in the current protection setting optimization model; (4) using a constrained multi-objective particle swarm algorithm to solve the objective function of the protection multi-objective setting optimization, and obtaining a group of Pareto optimal solutions of the objective function of the relay protection multi-objective setting optimization; and (5) using a K-means clustering algorithm to cluster the Pareto optimal solutions obtained from the step (4), and then selecting an optimal solution meeting the expected target of the relay protection from the clustering centers. The method can realize the comprehensive optimization of the reliability, selectivity, speediness and sensitivity of the relay protection.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of setting of relay protection, in particular to a current protection setting optimization strategy based on a multi-objective particle swarm algorithm with constraints. TECHNICAL BACKGROUND

[0002] With the economic development and social progress, the degree of electrification is increasingly high, and the residential power load is increasing. In order to ensure that the power demand of residents is met as soon as possible, the new line is often connected to the main network nearby, which solves the problem in a short time, but the change of the topology structure of the distribution network over a long period of time brings great influence to the protection of the distribution network. Especially when there are multiple levels of lines in the distribution network, the distance of each line is short, which easily leads to difficulties in setting the current protection. In GB / T 14285-2016, Technical Regulations for Relay Protection, only the necessary requirements for the line without electric reactance from the busbar of the power plant are made in 4.4.2.1, and the protection configuration and matching mode of each level of switch on the feeder are not specified, so in practice, various protection configurations and setting schemes are not reasonable. Unreasonable relay protection setting schemes lead to the problems of non-selective action and mismatch of protection, which are increasingly prominent, and are not conducive to the safe and stable operation of the distribution network and the reliable and sustainable power supply of the load.

[0003] For the three-section current protection method, the traditional relay protection of the distribution network has the advantages of simple principle, easy understanding and high reliability. Generally, when a fault occurs in the power supply system, it can be handled by cutting off, but this method cannot play its own role in the actual protection process of the distribution network due to the influence of factors such as the power grid connection mode and the system operation mode, and the problems of non-selective action and mismatch of protection are increasingly prominent, which has a great impact on the safe and stable operation of the distribution network and the reliable and sustainable power supply of the load. The current standard does not make provisions for this special case, which brings many inconveniences to the on-site staff. Even if the protection setting value is obtained through a large number of calculations and comprehensive on-site experience, it may not achieve the desired effect after being put into production.

[0004] If a relay protection setting optimization scheme can be established, the scheme can find the most suitable relay protection setting value for the changed primary structure, and the relay protection "four properties" of reliability, rapidity, sensitivity and selectivity can be used as comprehensive multi-objectives, which can solve the protection mismatch problem caused by the change of the topology structure, and is more conducive to the efficient work of the staff. SUMMARY

[0005] In order to solve the problems in the prior art, the present application discloses a current protection setting optimization strategy based on a multi-objective particle swarm algorithm with constraints, which can effectively achieve multiple optimization objectives of setting relay protection, and make the optimization objectives of the set relay protection have the characteristics of optimal comprehensive performance of relay protection in the power grid.

[0006] The application adopts the technical scheme of a current protection setting optimization strategy based on a constraint-containing multi-objective particle swarm algorithm, which comprises the following steps:

[0007] Step (1) establishing a target function of current protection setting optimization;

[0008] According to the "four properties" requirement of relay protection, the optimization indexes of relay protection are quantified, the target function of relay protection multi-objective optimization setting is established, and a relay protection multi-objective optimization setting model is obtained;

[0009] Step (2) determining the setting variables of the current protection setting optimization model: time setting value and current setting value;

[0010] Step (3) setting the constraint conditions of the setting variables of the current protection setting optimization model according to the mutual cooperation relationship between relays:

[0011] Constraint condition 1: upper and lower limit constraint of variables and parameters.

[0012]

[0013] Constraint condition 2: cooperation of current setting values of I, II and III sections of the same circuit breaker.

[0014] I 1i >I 2i >I 3i

[0015] Constraint condition 3: cooperation of current setting values between different circuit breakers.

[0016] I 1i,2i,3i >I 1i',2i',3i'

[0017] In the formula, i and i' represent the circuit breaker numbers at different positions on the same line, wherein i is closer to the system power source in position.

[0018] Constraint condition 4: cooperation of time setting values between the same circuit breaker.

[0019] T 3i >T 2i >T 1i

[0020] Constraint condition 5: cooperation of time setting values between different circuit breakers.

[0021]

[0022] In the formula, ΔT represents the time difference value of the circuit breakers at different positions on the same line, and is generally taken as 0.15-0.5s, and is taken as 0.2s in this paper.

[0023] Step (4) solves the objective function of the multi-objective setting optimization of the relay protection according to the constraint condition set in step (3) by using a multi-objective backbone particle swarm algorithm with constraints to obtain a group of Pareto optimal solutions of the objective function of the multi-objective setting optimization of the relay protection;

[0024] Step (5) clusters the group of Pareto optimal solutions obtained in step (4) by using a K-means clustering algorithm to obtain several representative categories, and then selects an optimal solution from the clustering centers that meets the expected target of the relay protection;

[0025] The relay protection setting optimization model in step (1) includes a speediness objective function of the relay protection setting, a sensitivity objective function of the relay protection setting and a reliability objective function of power supply.

[0026] The speediness objective function f1 of the relay protection setting is:

[0027]

[0028] In the formula, n is the total number of circuit breakers in the network. 1ij , T 2ij , T 3ij are the action times of the circuit breakers I, II and III under the jth short circuit, respectively, where i is the number of the circuit breaker, and i takes a value of 1-n.

[0029] The sensitivity objective function f2 of the relay protection setting is:

[0030]

[0031] In the formula, Ksen 1ij , Ksen 2ij , Ksen 3ij are the sensitivities of the circuit breakers I, II and III under a single short circuit, respectively, where i is the number of the circuit breaker, and i takes a value of 1-n.

[0032] The reliability objective function f3 of the power supply of the relay protection is:

[0033]

[0034] In the formula, P ij represents the probability of the misoperation and the refusal of the circuit breakers; S ij represents the power outage area caused by the error operation of the circuit breakers.

[0035] The optimization index of the relay protection in step (1) includes the sensitivity, the speediness and the reliability of the relay protection.

[0036] The multi-objective optimization setting model in step (1) is:

[0037] MinF(x) = [f1(x), f2(x), …, f M (x)]

[0038] In the formula, F is a target function, x is a position vector in a D-dimensional search space, f1, f2, …, f M are a group of functions mapping the D-dimensional search space to an M-dimensional target space. The smaller the obtained target function value is, the better the corresponding protection performance represents.

[0039] The specific process of solving the relay protection setting optimization target function by using the multi-objective backbone particle swarm algorithm with constraints in step (4) includes:

[0040] Step (4.1): Set the parameters required by the algorithm, including the particle swarm size N, the capacities Na and Na' of the feasible reserve set and the infeasible reserve set, the algorithm termination generation Tmax, and the mutation parameter a.

[0041] Step (4.2): t = 0, and the initial size of the particle swarm is N. Randomly assign an initial position to each particle in the particle swarm in the given feasible region; set the individual leader of each particle to itself; and set the feasible reserve set and the infeasible reserve set to empty sets.

[0042] Step (4.3): Calculate the fitness values of each particle in the particle swarm, and calculate their constraint violation degrees.

[0043] Step (4.4): Divide the particles into two categories of feasible solutions and infeasible solutions, and update the feasible reserve set and the infeasible reserve set in turn.

[0044] Step (4.5): Judge whether the termination criterion is met, that is, whether the termination generation is reached, and if the termination condition is met, the algorithm stops.

[0045] Step (4.6): For each particle in the particle swarm, select the global leader in turn, update the individual leader, generate a new particle position, and perform time-varying mutation.

[0046] The specific process of step (5) includes:

[0047] Step (5.1): Take each group of target function values of the relay protection as initial sample points, and perform data preprocessing;

[0048] Step (5.2): Cluster k from 2 to 9, and calculate the contour coefficient. The closer the contour coefficient is to 1, the better the clustering effect is, and the k value at this time is recorded.

[0049] Step (5.3): Select k sample points as initial division centers according to given k value;

[0050] Step (5.4): Calculate the distance of all sample points to each cluster center, and divide all sample points to the nearest cluster center;

[0051] Step (5.5): Calculate the average value of sample points in each division, and take it as a new center;

[0052] Step (5.6): Repeat step (5.4)~step (5.5) until the maximum iteration number is reached, or the change of cluster center is less than a predefined threshold.

[0053] Finally, according to the divided cluster center, a set of setting values closest to the actual needs is selected.

[0054] The technical scheme provided by the application has the beneficial effects that:

[0055] (1) The method is a setting scheme considering the overall performance optimization of all protections in the power grid, and the scheme can reflect the comprehensive optimization of the speed, sensitivity and power supply reliability of the relay protection.

[0056] (2) The method establishes multiple objective functions representing the performance of the relay protection, and constraint conditions describing the protection coordination relationship and other hardware factors, and adopts the multi-objective backbone particle swarm optimization algorithm integrating non-dominated sorting, elite reservation and time-varying variation to obtain the optimal solution set, so that different requirements of the power grid structure and load properties can be conveniently and quantitatively considered, and the global optimal setting scheme with different emphasis targets can be selected.

[0057] (3) The method has low investment cost, high intelligent level, wide engineering application prospect, and can be effectively promoted to the relay protection setting problem based on the overall performance optimization of all protections in the whole network. BRIEF DESCRIPTION OF DRAWINGS

[0058] The application will be further described below with reference to the drawings:

[0059] Figure 1 is a flow chart of the application;

[0060] Figure 2 is a schematic diagram of a distribution network structure; DETAILED DESCRIPTION

[0061] In order to better understand the purpose, technical scheme and technical effects of the application, the application will be further described below with reference to the drawings.

[0062] The application provides a current protection setting optimization strategy based on a multi-objective particle swarm algorithm with constraints, and the application is characterized in that Figure 1For the flow chart of the present application, the implementation process includes the following detailed steps.

[0063] Step (1): establishing a target function for current protection setting optimization;

[0064] According to the "four properties" requirement of relay protection, the optimization index of relay protection is quantified, a target function for relay protection multi-objective optimization setting is established, and a relay protection multi-objective optimization setting model is obtained;

[0065] The relay protection setting optimization model in step (1) includes a speediness target function, a sensitivity target function and a reliability target function for power supply of relay protection setting;

[0066] The speediness target function f1 for relay protection setting is:

[0067]

[0068] In the formula, n is the total number of circuit breakers in the network. 1ij , T 2ij , T 3ij respectively the action time of the circuit breaker I, II, III section protection under the jth short circuit, wherein i is the number of circuit breakers, and its value is 1-n.

[0069] The sensitivity target function f2 for relay protection setting is:

[0070]

[0071] In the formula, Ksen 1ij , Ksen 2ij , Ksen 3ij respectively the sensitivity of the circuit breaker I, II, III section protection under single short circuit, wherein i is the number of circuit breakers, and its value is 1-n.

[0072] The power supply reliability target function f3 for relay protection is:

[0073]

[0074] In the formula, P ij represents the probability of circuit breaker malfunction and refusal; S ij represents the power outage area caused by the error action of the circuit breaker. The relay protection multi-objective optimization setting model in step (1) is:

[0075] Min F(x) = [f1(x), f2(x), …, f M (x)]

[0076] In the formula, F is the target function; x is the position vector in the D-dimensional search space; f1, f2, …, fM is a set of functions that map the D-dimensional search space to the M-dimensional target space. The smaller the resulting target function value, the better the corresponding protection performance represented.

[0077] Step (2): Determine the setting variables in the current protection setting optimization model: time setting and current setting;

[0078] Step (3): According to the mutual coordination relationship between relays, set the constraint conditions of the setting variables in the current protection setting optimization model:

[0079] Constraint condition 1: Upper and lower limit constraints of variables and parameters.

[0080]

[0081] Constraint condition 2: Coordination of current settings of I, II, and III sections of the same circuit breaker.

[0082] I 1i > I 2i > I 3i

[0083] Constraint condition 3: Coordination of current settings between different circuit breakers.

[0084] I 1i,2i,3i > I 1i',2i',3i'

[0085] In the formula: i, i' represent the circuit breaker numbers at different positions on the same line, where i is closer to the system power source in position.

[0086] Constraint condition 4: Coordination of time settings between the same circuit breaker.

[0087] T 3i > T 2i > T 1i

[0088] Constraint condition 5: Coordination of time settings between different circuit breakers.

[0089]

[0090] In the formula: ΔT represents the time difference value of circuit breakers at different positions on the same line, generally taking a value between 0.15-0.5s, and the value taken in this paper is 0.2s.

[0091] Step (4): According to the constraint conditions set in step (3), use the multi-objective backbone particle swarm algorithm with constraints to solve the target function of relay protection multi-objective setting optimization, and obtain a set of Pareto optimal solutions of the target function of relay protection multi-objective setting optimization;

[0092] The specific process of solving the relay protection setting optimization objective function in the step (4) includes:

[0093] Step (4.1): setting the parameters required by the algorithm, including the particle swarm size N, the capacities Na and Na' of the feasible reserve set and the infeasible reserve set, the algorithm termination generation Tmax and the mutation parameter a, etc.

[0094] Step (4.2): t=0, the initial size of the particle swarm is N. Randomly assign an initial position to each particle in the particle swarm in the given feasible region; set the individual leader of each particle to itself; set the feasible reserve set and the infeasible reserve set to empty sets;

[0095] Step (4.3): calculate the fitness value of each particle in the particle swarm and calculate their constraint violation degree;

[0096] Step (4.4): divide the particles into two categories of feasible solutions and infeasible solutions, and update the feasible reserve set and the infeasible reserve set in turn;

[0097] Step (4.5): judge whether the termination criterion is met, that is, whether the termination generation is reached, if the termination condition is met, the algorithm stops;

[0098] Step (4.6): for each particle in the particle swarm, select the global leader in turn, update the individual leader, generate a new particle position, and perform time-varying mutation.

[0099] Step (5): using the K-means clustering algorithm to cluster the representative several categories from the set of Pareto optimal solutions obtained in step (4), and then selecting an optimal solution from the clustering centers that meets the expected target of the relay protection;

[0100] The specific process of the step (5) includes:

[0101] Step (5.1): taking each set of objective function values of the relay protection as initial sample points, and performing data preprocessing;

[0102] Step (5.2): clustering k from 2 to 9, and calculating the silhouette coefficient, the clustering effect is better when the silhouette coefficient is closer to 1, and record the k value at this time;

[0103] Step (5.3): according to the given k value, select k sample points as the initial division center;

[0104] Step (5.4): calculate the distance of all sample points to each clustering center, and divide all sample points to the nearest clustering center;

[0105] Step (5.5): Calculate the average value of the sample points in each division, and take it as a new center;

[0106] Step (5.6): Repeat steps (5.4) to (5.5) until the maximum number of iterations is reached, or the change of the clustering center is less than a certain predefined threshold.

[0107] Finally, according to the divided clustering center, a set of setting values closest to the actual needs is selected.

[0108] To verify the correctness and effectiveness of the method for relay protection setting optimization proposed in the present application, the method is used to perform multi-objective setting optimization of relay protection for the distribution network shown in Fig. 1, and the network parameters are shown in Table 1, and the setting results are shown in Table 2. Figure 2

[0109] Table 1 Node branch parameters

[0110]

[0111] After the optimization calculation, the obtained setting results are as follows:

[0112] Table 2 Setting calculation results

[0113]

[0114] As can be seen from Table 1 and Table 2, the method of the present application can conveniently and quantitatively consider different requirements of power grid structure and load properties, experience and preference of relay protection engineers, and select current protection global optimal setting schemes with different emphasis targets. The method has low investment cost, high intelligent level, wide engineering application prospect, and can be effectively promoted to the relay protection setting problem based on the overall performance optimization of all protections in the whole network.

[0115] Although the specific embodiments of the present application are described above in combination with the drawings, the present application is not limited to the scope of the embodiments, and those skilled in the art should understand that various modifications or changes made to the technical solutions of the present application without creative labor are still within the protection scope of the present application.​

Claims

1. A current protection setting optimization strategy based on a constrained multi-objective particle swarm optimization algorithm, characterized in that, Includes the following steps: (1) Establish the objective function for current protection setting optimization; Based on the "four characteristics" requirements of relay protection, the optimization indicators of relay protection are quantified, the objective function of multi-objective optimization setting of relay protection is established, and the multi-objective optimization setting model of relay protection is obtained. (2) Determine the setting variables in the current protection setting optimization model: time setpoint and current setpoint; (3) Based on the coordination relationship between relays, set the constraints on the setting variables in the current protection setting optimization model: Constraint 1: Upper and lower bound constraints for variables and parameters: Constraint 2: Coordination of current settings for sections I, II, and III of the same circuit breaker: I 1i >I 2i >I 3i Constraint 3: Coordination of current settings between different circuit breakers: I 1i,2i,3i >I 1i',2i',3i' In the formula: i, i' represent the circuit breaker numbers at different locations on the same line, where i is closer to the system power source in terms of location; Constraint 4: Coordination of time settings between the same circuit breakers: T 3i >T 2i >T 1i Constraint 5: Coordination of time settings between different circuit breakers: In the formula: ΔT represents the time difference between circuit breakers at different locations on the same line, which is generally between 0.15 and 0.5s, and is 0.2s in this paper; (4) Based on the constraints set in step (3), the objective function of the multi-objective setting optimization of relay protection is solved by using the constrained multi-objective backbone particle swarm algorithm, and a set of Pareto optimal solutions of the objective function of the multi-objective setting optimization of relay protection is obtained. (5) The K-means clustering algorithm is used to cluster a set of Pareto optimal solutions obtained in step (4) to obtain several representative classes, and then an optimal solution that meets the expected goal of relay protection is selected from the cluster centers.

2. The current protection setting optimization strategy based on a constrained multi-objective particle swarm optimization algorithm as described in claim 1, characterized in that, The optimization indicators of relay protection in step (1) include the sensitivity, speed and power supply reliability of relay protection. The relay protection setting optimization model in step (1) includes the speed objective function, sensitivity objective function and power supply reliability objective function of relay protection setting; The objective function f1 for the speed-up setting of the relay protection is: In the formula: n is the total number of network circuit breakers; T 1ij T 2ij T 3ij The operating times of the protection stages I, II, and III of the circuit breaker under the j-th short circuit are respectively, where i is the circuit breaker number, and its value ranges from 1 to n; The sensitivity objective function f2 of the relay protection setting is: In the formula: Ksen 1ij Ksen 2ij Ksen 3ij These are the sensitivities of the circuit breaker's I, II, and III protection stages under a single short circuit, respectively, where i is the circuit breaker number, and its value ranges from 1 to n; The power supply reliability objective function f3 for the relay protection is: In the formula: P ij S represents the probability of a circuit breaker malfunctioning or failing to operate; ij This indicates the area of ​​power outage caused by a circuit breaker malfunction.

3. The current protection setting optimization strategy based on a constrained multi-objective particle swarm optimization algorithm as described in claim 1, characterized in that, The multi-objective optimization setting model for relay protection in step (1) is as follows: MinF(x)=[f1(x),f2(x),…,f M (x)] In the formula: F is the objective function; x is the position vector in the D-dimensional search space; f1, f2, ..., f M It is a set of functions that map the D-dimensional search space to the M-dimensional target space; the smaller the value of the target function, the better the corresponding protection performance.

4. The current protection setting optimization strategy based on a constrained multi-objective particle swarm optimization algorithm as described in claim 1, characterized in that, The specific process of solving the relay protection setting optimization objective function using the constrained multi-objective backbone particle swarm algorithm in step (4) includes: Step (4.1): Set the parameters required for the algorithm, including the particle swarm size N, the capacities Na and Na' of the feasible and infeasible storage sets, the algorithm termination generation Tmax, and the mutation parameter a; Step (4.2): t = 0, initial particle swarm size N; randomly assign an initial position to each particle in the particle swarm within a given feasible region; set the individual guide of each particle to itself; set both feasible and infeasible reserve sets to empty sets; Step (4.3): Calculate the fitness value of each particle in the particle swarm and calculate their constraint violation degree; Step (4.4): Divide the particles into two categories: feasible solutions and infeasible solutions, and update the feasible reserve set and infeasible reserve set in turn; Step (4.5): Determine whether the termination criterion is met, i.e. whether the termination algebra has been reached. If the termination condition is met, the algorithm stops. Step (4.6): For each particle in the particle swarm, select a global leader in sequence, update the individual leader, generate a new particle position, and perform time-varying mutation.

5. The current protection setting optimization strategy based on a constrained multi-objective particle swarm optimization algorithm as described in claim 1, characterized in that, The specific process of step (5) includes: Step (5.1): Use each set of objective function values ​​of the relay protection as initial sample points for data preprocessing; Step (5.2): Cluster k from 2 to 9 and calculate the silhouette coefficient. The closer the silhouette coefficient is to 1, the better the clustering effect. Record the value of k at this time. Step (5.3): Select k sample points as the initial partition centers based on the given k value; Step (5.4): Calculate the distance from all sample points to each cluster center, and assign all sample points to the nearest cluster center; Step (5.5): Calculate the average value of the sample points in each partition and use it as the new center; Step (5.6): Repeat steps (5.4) to (5.5) until the maximum number of iterations is reached, or the change in cluster centers is less than a certain predefined threshold; Finally, based on the divided cluster centers, the set of tuning values ​​that best matches the actual needs is selected.

Images